Automated chest x-ray analysis Page: 2 of 10
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to UNT Digital Library by the UNT Libraries Government Documents Department.
Extracted Text
The following text was automatically extracted from the image on this page using optical character recognition software:
AUTOfIlATED CHEST X-RAY ANALYSIS'
E. L. Hall
B. K. Rougc
R. P. Krugcr"
Depr.rtments of Flec:ri3l Lnrinecring & Radiology
Ur.iversity of vu:frvrn California
Los Angeles, California
*Los Alamos Scientific Laboratory
Los Alamos, New Mexico
Abstract
A computer hardware-software system has beer, developed for extracting optical and digital mea-
surements from chest x-ravs and usin; these ieac":remc:2ns to clarsifv the severity of lung disease. The
system consists of a dici:ar imare scanner and an cp:ical sys:em consisting of an x'- y controller for film
transportation and an RE: Fra,;nhofer difraction pa:.tcrzn samnltng unit. Ltitially the classification has been
of opacity texturali nau:-rs associated with coal workers pneumoconiosis (CWP). However, the system
appears to be ex-andable to a:1 diseases descriced in tie international standard which include lung disease,
heart disease, and car.cer witn only software moi'ication. Recently extensions to large capacity cetection,
measures of lung vasculcrity and lung and heart volume have been made. Results indicate classification
accuracy rates for pneumoconioses comparable to visual rcadinrs of the films by expert radiologists, and
significantly better than non-radiologist physicians who often read these films.
Introduction
The successful use of computers in many diverse applications depends largely upon the development
of a successful software system. The purpose of this paper is to describe recent developments in a soic-
ware system called DR. ADA. (Diagnostic Radiology Aid to Diagnosis by Automatic Measurements) which
was developed to provide measure nts anu aecisions"srom cnest x-ray films
Previous studies on the Dr. Adam system have been reported in detail't, 2, 3). This paper will
report recent results and updates while explaining the system in general. References will be made to the
in-depth discussions of algorithmic justificationL Studies are currently in progress for tne detection of
coal workers pneumodcniosis (CWP), pulmonary ederc, and abnormal heart and lung volume. This paper
will discuss the results of the CWP studies. Other capabilities will be reported at seminar 9 of session 4
of the SPIE symposiunn by Dr. A. Franklin Turner.
The extent of disease of CWP is judged on 12 mnor and 4 major categories which are based on the
profusion of small regular and irregular opacities per unit area. A fifth major profusion category consists
of still larger, regular and irregular shaped lesions associated -ith progressive massive fibrosis (PAIF).
This fifth category often occurs when smaller opacities coalesce.
To determine the overall film profusion category, a decision as to the erent of the opacities in the
lung field must first be made by the physician. Each lung field is vertically divided into three -ones. Over-
all film profusion category is established by visually averaging the profusion in the affected zones. The
result is a visually derived indication of zonal extent, opacity type, and profusion category for each film.
Each film is read by at least two govern.=en certified physicians according to the International
Labor Organization (I LO) 1971 Classification of Radiograpns o, the Pneumoconiosist1. Depending on the
'agreement of these initial readers, a film decision could require a panel of readers. It is anticipated that
the Dr. Adam system could assist in handling the mounting quantities of CWP x-rays being taken. Specifi-
cally. this automated system could screen out numerous normal films and indicate the abnormal ones wt.ich
require a physicians interpretation. As mentioned previously, the measurements could reflect abnora ali-
ties of many types other than pneumoconiosis. For example, lung vascularity measurements can be made.
Abnoemal vascularity conditions are indicative of both pulmonary edema and cancer.
Prototype System
Using experience gained from a continually growing computer system, the final configuration is
being trained as the Dr. Adam prototype system. Originally, dtta acquisition was implemented on a
Hewlett Packard 2100 which cor:rols the Dicomed imnare digitizer, the Recognition Systems optical spec-
trum analyzer, the x - y film transport and storage peripherals. The data obtained was analysed by using
programs on an LOM 370 and a PDP 10. The changeover to a stand alone Dr. Adam system began with the
acquisition of a PDPl process controller. Additional hardware and software was incorporated and modi-
ied to provide greater speed and data handling capabilities. After training is completed, the system will
be independent with data acquisition, storage and display.
While the initial prototype system consists of physically distinct but electronically integrated sub-
systems. this discussion will present a feasible scenario for their physical integration. A block diagram
is shown in Figure 1.
V This work performed under NIOSH Contract iSM 99-73-92
4&, /97u Si,-c CoMCtV M" I
SRA-. D/tuo, CI/c
Upcoming Pages
Here’s what’s next.
Search Inside
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Hall, E. L.; Rouge, B. K. & Kruger, R. P. Automated chest x-ray analysis, article, January 1, 1977; New Mexico. (https://digital.library.unt.edu/ark:/67531/metadc1443532/m1/2/: accessed April 18, 2024), University of North Texas Libraries, UNT Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.